As is well understood, with any microfluidic assay or capture system there is the potential for failure, and with complex systems, typically, there are numerous potential failure modes. It is greatly desired to provide techniques that, in concert, avoid or minimize these problems.
Examples of failure modes for microfluidic assay systems relate to flows in microfluidic channels and to valves and pistons that control the flows according to a pre-determined assay protocol. The failure modes occur with any microfluidic system, but can be of particular concern when the assay is performed within a microfluidic cartridge. Blockage of a microfluidic channel and inability of a valve to open or close are examples of failure. If a valve does not open, flow is prevented; if it does not close completely, valve leakage may occur at an inopportune time. There is also the possibility of a contaminant in the microfluidic channels.
There are also potential for human errors. For example, most samples for immunoassays, e.g., human plasma or human serum samples, are diluted with a diluent at a prescribed ratio, for example one-to-one (one part sample to one part diluent) or one to five. It is important for the proper ratio to be supplied, but personnel may improperly prepare the samples.
With microfluidic assays in general, and especially automated microfluidic immunoassays performed within portable cartridges, there are many steps in the assay protocol that need to occur with specific timing and specific reagents.
It is necessary to minimize the risks of error and, as well, to know when error occurs. For instance it is necessary to know when a buffer liquid or reagent is being flowed through a microfluidic channel and when a sample is being flowed, and whether, in each case, it is flowed at the proper rate and/or for the proper duration. It is also necessary to know whether there is leakage or flow into regions where no flow can be permitted. Further, when a liquid volume is displaced in a pulsed flow type microfluidic system, for instance as a reciprocating piston pump pushes small slug (portions) of liquid sequentially through a channel, it is important to have a precisely determined quantity of fluid in each slug.
Problems in attempting to obtain this information arise. For instance, because liquids flow in microfluidic channels are very tiny (e.g., 100-200 microns cross section width and depth, only millimeters in length) flows are difficult to visualize. The channels are so small that it is difficult for the human eye to observe the fact that liquid is not flowing where it is desired. The assay reagents are typically transparent, compounding the difficulty of visual or optical observation.
The problems are especially acute when seeking highly accurate quantification in a microfluidic assay. In quantifying assays it is desired that a given amount of immobilized capture agent be exposed to a given amount of various fluids to enable reactions over defined times so that results can be compared to a standard to enable the quantification. Results need to be determined with an overall coefficient of variation of less than 10% (accuracy within 10%), preferably much less.
Thus, to be quantitative, assays require consistent run-to-run performance. For example in an assay employing a fluorescent dye conjugated with immobilized, captured moieties, the concentration and the volume of the buffer or wash liquid, of secondary reagents such as antibodies, and of fluorescent dye all need to be the same from run to run if one is to compare the result to a standard calibrating curve precisely generated from previous calibrating runs. This is particularly true in blood testing in which a patient human plasma or serum sample is measured for the presences or the quantity of specific health-related analytes, for instance, antibodies such as interleukins (a class of antibodies called cytokines), e.g., IL5 or IL6. There are many other classes of antibodies to be measured in plasma or serum. Similarly there are many DNA and other nucleic acid determinations that are desired in the drug development, medical and other fields, including public health, agriculture and food processing.
For these reasons it is important to verify that at the end of an assay when a result is generated, that the result has been produced precisely according to the desired protocol.
The result of an assay is typically measured by detecting an emanation, e.g., a fluorescence intensity, from a reaction site. The emanation may come from a bead, a micro particle or an immobilized spot. As presently preferred, it comes from an immobilized glass nano-reactor (GNR) in the form of a small hollow tube or micro-length tube, of length no more than 1000 micron, typically less than 500 micron, with capture agent, e.g., antibody, immobilized on its inside surface.
The fluorescence intensity from the region of the capture agent is essentially all that is measured at the completion of many assays. That fluorescence intensity is compared to a calibration curve. From the calibration curve the unknown concentration of the analyte is determined. For the calibration curve to be valid to a particular run, it is necessary that all of the conditions for that assay are repeated specifically and reproducibly from run to run. Improved means to measure such conditions are to be greatly desired.
For the following description of novel techniques for performing and monitoring microfluidic assays, it is important that the exact location of features on a microfluidic cassette be known. Novel techniques for doing this are described later herein.